2 prt* party variables

Information on ESS party IDs from “party-voted-for” (prtc*) and “party-close-to” (prtc*) questions – see also section “ESS party data structure” in manuscript.

Code

library(conflicted)

library(tidyverse)
conflicts_prefer(dplyr::filter, .quiet = TRUE)

library(knitr)
library(reactable)

Code

# ESS datasets in rds and with selected variables created in "01-ess-prt.R"
ess9_raw <- read_rds("data/ess-waves/ESS9e03_1.rds")
ess_raw <- read_rds("data/02-ess-select.rds")

# long format ESS prt* variables
prt_raw <- read_rds("data/01-ess-prt.rds")

# prtv* (single per country) and prtc* ESS variable with ESS numeric ID
prtvc_raw <- read_rds("data/01-ess-prtv-prtc.rds")

2.1 prt* variables

All ESS rounds include two types of survey questions with party information.

prtv* — “Party voted for in last national election, [country]?”
prtc* — “Which party feel closer to, [country]?”

Code

ess_raw |>
  summarise(
    n = n(),
    n_countries = n_distinct(cntry),
    n_prtv = n_distinct(prtv),
    n_prtc = n_distinct(prtc),
    .by = "essround"
  ) |>
  arrange(essround)

essround	n	n_countries	n_prtv	n_prtc
1	42359	22	251	248
2	47537	25	281	274
3	43000	23	248	251
4	56752	29	309	296
5	52458	27	321	302
6	54673	29	331	331
7	40185	21	258	252
8	44387	23	283	271
9	49519	29	363	374
10	58810	30	380	390

2.2 ESS-9 example

We use the ESS-9 integrated file to describe the structure of the prt* variables in ESS data files.

Each ESS round uses country level variables for the prt* variables (e.g. prtvtcat — party-voted-for Austria ESS-9).

These prt* variables include the following elements:

starting with prt
indicating the type of prt variable
- v — “party-voted-for”
- c — “party-close-to”
two character country code
electoral tier number for Germany and Lithuania (prtv* only)

Code

ess9_prt <-
  ess9_raw |>
  select(cntry, starts_with(c("prtv", "prtc"))) |>
  pivot_longer(!cntry, names_to = "variable", values_to = "party") |>
  na.omit()

ess9_prt_n <-
  ess9_prt |>
  summarise(
    cntry = first(cntry),
    responses = n(),
    parties = paste(unique(party) |> as.character() |> sort(), collapse = ", "),
    .by = c(cntry, variable)
  )

ESS-9 prt* variables by country

Code

tbl_out <-
  ess9_prt_n |>
  summarise(
    n = n(),
    variables = paste(variable |> sort(), collapse = ", "),
    .by = cntry
  )

if (knitr::is_html_output()) {
  tbl_out |>
    reactable(searchable = TRUE, striped = TRUE)
} else {
  tbl_out |>
    slice((1:5))
}

Germany and Lithuania include multiple prtv* variables asking for voting decisions in each electoral tier. These variables include a number for the tier in the variable name.

We use the national tier (“prtvede2”, “prtvblt1”) as the primary “party-voted-for” variable.

Code

ess9_prt |>
  filter(str_detect(variable, "\\d")) |>
  distinct(cntry, variable) |>
  arrange(cntry, variable)

cntry	variable
DE	prtvede1
DE	prtvede2
LT	prtvblt1
LT	prtvblt2
LT	prtvblt3

2.3 prtv* ID differences

ESS may use different IDs across ESS rounds

Code

prtv_ids <-
  prt_raw |>
  mutate(party = as.character(party)) |>
  filter(
    str_detect(variable, "^prtv"), # party-voted-for variable
    !str_detect(variable, "de1|lt[23]") # national vote only DEU and LTU
  ) |>
  select(-idno, -variable, -ess_id) |>
  distinct() |>
  arrange(party_id) |>
  pivot_wider(
    names_from = party_id,
    values_from = party,
    names_prefix = "id_"
  ) |>
  arrange(cntry, essround)

e.g. Netherlands prtv* rounds 1–10

Code

prtv_ids |>
  filter(cntry == "NL") |>
  select(essround, id_4:id_6)

essround	id_4	id_5	id_6
1	List Pim Fortuyn	Democrats ’66	Green Left
2	List Pim Fortuyn	Democrats ’66	Green Left
3	List Pim Fortuyn	Democrats ’66	Green Left
4	List Pim Fortuyn	Democrats ’66	Green Left
5	Christian Democratic Appeal	Socialist Party	Democrats ’66
6	Christian Democratic Appeal	Socialist Party	Democrats ’66
7	Socialist Party	Christian Democratic Appeal	Democrats ’66
8	Socialist Party	Christian Democratic Appeal	Democrats ’66
9	Socialist Party	Christian Democratic Appeal	Democrats ’66
10	Socialist Party	Christian Democratic Appeal	Democrats ’66

2.4 prtv/prtc ID differences

ESS party IDs may differ between the prtv* and prtc* variables.

Code

tbl <-
  prtvc_raw |>
  mutate(
    prtv_party_id = str_extract(prtv, "\\d+-\\d+") |> str_remove("\\d+-") |> as.integer(),
    prtc_party_id = str_extract(prtc, "\\d+-\\d+") |> str_remove("\\d+-") |> as.integer()
  ) |>
  filter(prtv_party_id == prtc_party_id & prtv_party != prtc_party) |>
  select(cntry, essround, party_id = prtv_party_id, prtv_party, prtc_party) |>
  distinct() |>
  arrange(essround, cntry, prtv_party)

Examples from six countries in ESS-9

Code

tbl |>
  filter(essround == 9) |>
  slice(2, 5, 18, 27, 29, 38)

cntry	essround	party_id	prtv_party	prtc_party
BG	9	2	Balgarska sotsialisticheska partiya (BSP)	Dvizhenie za prava i svobodi (DPS)
FI	9	10	Green League	Independence Party
LT	9	2	Homeland Union - Lithuanian Christian Democrats (TS-LKD)	Lithuanian Peasant and Greens Union (LVZS)
PL	9	3	Nowoczesna	Platforma Obywatelska
PT	9	17	Votou em branco/ nulo	CDS-PP
SK	9	5	ĽS Naše Slovensko	Christian Democratic Movement (KDH)

Code

tbl_v <-
  prtvc_raw |>
  filter(str_detect(prtv, "FI-(8|9|10)-10-v")) |>
  distinct(essround, prtv, prtv_party)

tbl_c <-
  prtvc_raw |>
  filter(str_detect(prtc, "FI-(8|9|10)-10-c")) |>
  distinct(essround, prtc, prtc_party)

tbl_out <-
  tbl_v |>
  left_join(tbl_c)

write_csv(tbl_out, "figures-tables/table-1_ess-incoherence.csv")

2.1 prt* variables

2.2 ESS-9 example

2.3 prtv* ID differences

2.4 prtv*/prtc* ID differences

2.4 prtv/prtc ID differences